Freesound Forums

Support for uploading in wavpack format is something that could be considered, also because we have some plans to have better support for ambisonics recordings. I’ll add that to our list and see how hard it would be to implement support for it.

That would be awesome, thank you!

Hi everyone,

I saw that adding WavPack support was quite easy in our infrastructure so I just added it. Please try it and let me know if you find any problems... I did some quick tests and it worked fine but sounds were rather simple and only stereo.

Wow this was quick, amazing!
I just uploaded 3 different recordings in 16 channel Ambisonics B-Format as wavepack, and it seems to work perfectly!
I was only wondering about the waveform of the preview, which looks rather small despite one of the recordings peaking at -1 dB (overall -28 LUFS).
This brings me also another question in regard to which channel you're using for the preview, or are you already doing some sort of conversion to binaural? Simply using the first channel would probably make the most sense, as it being the omni channel.

As I can tell by the great feedback I got when sharing my Ambisonics recordings in one of the Ambisonics Facebook groups, the need for a good place for our B-Formats is really needed, as binaural is just not sufficient.
So thank you for your work, this is a big deal for every Ambisonics enthusiast who wants to upload recordings!

I found a small problem, it seems to be that the bulk description of audio files doesn't seem to recognize wavepack files yet. It tells me that there is an invalid file extension. Strangely, it worked on one of the files...

Hi @JakobGilie,

That is quite strange, specially if it did work once. Can you double check that the error still persists? I looked at the code and as far as I can tell .wv format should work normally. Note that we only support the .wv format, not the hybrid format which requires 2 audio files.

Just out of curiosity, is WavPack the most common format for exchanging Ambisonics recordings or is there another format which is more popular?

As far as I remember the processing for the sound images takes the first channel in a multi-channel setting. Preview files (mp3 and ogg) keep the first 2 channels if I'm not mistaken, but we don't do any ambisonics->binaural processing (also with the low quality of the preview maybe the binaural effect would be gone). We have plans to try to find a proper solution for this, so if we can identify the ambisonics format of the upload sound (maybe by requesting users to add a specific tag to the description or using some specific file format), then we could do the binaural mixdown and use that as a preview (and also increase a bit the quality of the preview to make sure some spatial information is preserved). We'll see if we can make that happen soon.

There was probably something wrong with my CSV file. I already described my recordings manually and will check if the problem persists next time.

As far as I know .wav is the only standard used in Ambisonics recordings and exchange (e.g. for competitions and recordings). The file size restriction on freesound started me thinking about different possible file formats, also to save some disk space. So wavepack seems to be the only possible solution, as FLAC is restricted to 8 channels, in my tests Reaper rendered only 6 channels in AAC and all the other option are not using lossless compression.

I think it should be sufficient to use the first channel of Ambisonics recordings as a preview. Even it being only mono, one can get a good impression of the recording before downloading Binaural conversion would be nice but probably not trivial as there are so many variables that influence the conversion process.